### Final Presentation of Pitt Group

```DTI group (Pitt)
Instructor: Kevin Chan
Kaitlyn Litcofsky & Toshiki Tazoe
7/12/2012
Aims
1. Understand basic principles of MRI
2. Examine factors affecting DWI
1. b-values
3. Examine effect of b-values on DTI
4. Integrate fMRI and DTI
MR signal
S = M0 ・ (1 - e -TR/T1) ・ (e -TE/T2)
Subject 1
TE: 146 ms
TE: 73 ms
TE: 18 ms
TR: 500 ms
TR: 6000 ms
MR signal and diffusion
S = M0 ・ (1 - e -TR/T1) ・ (e -TE/T2) ・ e-bD
= S0 ・ e-bD
D = diffusion coefficient
Gaussian distribution of diffusion
If DW signal comes from free diffusion, gradient magnetic pulse would decay
DW signal mono-exponentially with b-value
Diffusivity across b-value decreases linearly
Diffusion coefficient across b-value is constant
S = S0 ・ e-bD
b-value
D = ln(S/S0) / -b
D
Signal
ln(S/S0)
-bD = ln(S/S0)
b-value
b-value
Effect of b-values on DWI:
Water phantom
free diffusion?
(NiS04.6H20/NaCL)
Mean DWI (50 directions)
b=0
b=500
b=1000
b=1500
Signal, DWI
(mm2/s)
1200
300
ln S/S0
0
-0.5
-1
-1.5
-2
-2.5
-3
-3.5
-4
-4.5
1000
800
600
400
200
0
0
500
1000 1500 2000 2500
b-value (s/mm2)
SNR ≒ 1
SNR ≒ 1
b=2000
b=2500
0
(mm2/s)
0.0025
0.002
0.0015
0.001
0.0005
0
0
500
1000 1500 2000 2500
b-value (s/mm2)
500
1000
1500
2000
b-value (s/mm2)
2500
Effect of b-values on DWI:
free diffusion?
Mean DWI (50 directions)
b=0
b=500
(mm2/s)
250
DWI
200
b=1000
300
b=1500
ln(S/S0)
0
b=2000
0
b=2500
(mm2/s)
0.001
0.0008
-0.5
0.0006
150
-1
0.0004
100
-1.5
50
0
0.0002
0
-2
0
500
1000 1500 2000 2500
b-value (s/mm2)
0
500
1000 1500 2000 2500
b-value (s/mm2)
500
1000
1500
2000
2500
b-value (s/mm2)
WM
GM
Effect of varied b-values on DWI:
x
DW signal at diffusion gradient (0.79, 0.61, 0.06)
R
y
L
Fast diffusion Slow diffusion
b value (s/mm2)
Effect of varied b-values on DTI
Fractional anisotropy: FA
λ1
Axial diffusivity: λ//
z
(mm2/s)
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0.002
λ2
38%
0.0015
10%
0.001
17%
0.0005
25%
y
λ3
0
500
1000
1500
2000
2500
500
b-value (s/mm2)
(mm2/s)
0.0007
0.0009
0.0008
0.0007
0.0006
0.0005
0.0004
0.0003
0.0002
0.0001
0
21%
0.0006
0.0005
0.0004
0.0003
0.0002
30%
0.0001
0
1500
2000
2500
Mean diffusivity
(mm2/s)
1000
1500
b-value (s/mm2)
500
1000
2000
b-value (s/mm2)
2500
20%
WM
GM
37%
500
1000
1500
b-value (s/mm2)
2000
2500
x
Effect of varied b-values on DTI
Voxel-based method, 8 subjects
– Tract-based Spatial Statistics
– FA at b = 1000 s/mm2 and b = 2500 s/mm2
P = 0.05
P ≒ 0.00
DTI tractography
• Inputs
– Principal vector
– FA
• Tractography: FACT method
–
–
DTIStudio
Fiber Assignment by Continuous Tracking (FACT)
approach
•
•
Start/Stop tracking threshold: FA = 0.2
Turn threshold: 70 degrees
DTI tractography by manual ROI
Corticospinal tract
Slice 0
Slice 31
Number of voxels passed through
3000
2500
2000
1500
1000
500
0
500
1000
1500
2000
2500
b value (s/mm2)
b=500
b=1000
b=1500
b=2000
b=2500
fMRI data as DTI seed regions
•
Compare tractography of posterior visual pathways for upper and lower field
visual stimulation at b=1000 s/mm2 and b=2500 s/mm2
1.
–
Block design
•
•
–
–
–
2.
TR = 2000 ms
TE = 26 ms
8 subjects
fMRI analysis
–
3.
Rest-Upper-Rest-Lower
12 s blocks, 6 repetitions
FSL FEAT
Create masks for DTI from fMRI activation maps
Rest
Upper
field
stimulation
Lower
field
stimulation
fMRI data as DTI seed regions
b = 1000 s/mm2
Upper visual
field stimulation
b = 2500 s/mm2
Lower visual
field stimulation
fMRI data as DTI seed regions
n=8
# of voxels
Mean FA
1200
0.7
1000
0.6
0.5
800
0.4
600
0.3
400
0.2
200
0.1
0
0
Upper b=2500
visual field
b=1000
Lower visual
field
b=1000
b=2500
Upper visual field
b=1000
b=2500
Lower visual
field
b=1000
b=2500
Conclusions
1. Diffusion is not free / Gaussian-distributed in the brain
– b-values and direction of gradient affects DWI
– b-values affect DTI metrics
 Caution has to be taken when interpreting brain DWI/DTI metrics at
different b-values
2. Lower b-values (at ~1000 s/mm2 or 1/ADC) may be more
beneficial for evaluating DTI metrics given the higher SNR
and potentially smaller errors in estimation (Jones & Basser, 2004)
3. Higher b-values (e.g., 2500 s/mm2) may be more beneficial
for tractography given higher number of voxels traced, likely
as a result of greater sensitivity in detecting smaller fibers (Rane,
Nair & Duong, 2010)
Thank you!
Dr. Kevin Chan
Dr. Seong-Gi Kim
Dr. Bill Eddy
Tomika Cohen
Rebecca Clark
MNTP program
```